Acoustic models can help an electronic device interpret speech inputs, for instance, during implementation of speech-to-text systems. By way of example, acoustic models allow electronic devices to better interpret phonetic and prosodic aspects of speech, such as various specific words and phrases, accents, tones, and rhythms. Acoustic models intended for general use, however, are computationally demanding and impose a large memory requirement. Accordingly, less computationally capable devices typically cannot implement such acoustic models.